Composite Bird's Eye View Stop Line Detection
Solution Overview
Problem
Existing stop line detection systems in autonomous vehicles are unreliable due to dirt, tire polish, and illumination variations, leading to inaccurate edge detection and safety risks, and are complex, increasing processing time and reducing vehicle safety.
Innovation Solution
A system and method using a front view camera to generate composite bird's eye view images, which includes an original and rotated image, employing a fast line filtering algorithm to simultaneously extract stop lines and lane markers, utilizing convolution kernels and modified Hough transformations to detect white painted lines robustly and efficiently.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If known edge detectors (Sobel and Canny) are used to detect stop lines, then the detection can be performed, but the detection accuracy deteriorates due to dirt, illumination variation, and noise
Solution Approach 1:
The patent segments the stop line detection problem into multiple stages: candidate region identification based on lane marker geometry, line endpoint detection, and perpendicular line search. This segmentation allows each stage to focus on specific features, improving overall detection reliability while maintaining precision by filtering out dirt and noise at each stage.
Solution Approach 2:
The patent performs preliminary actions by first detecting lane markers and establishing geometric constraints before searching for stop lines. This preliminary geometry establishment creates a constrained search space that eliminates false detections from dirt and illumination variations, thereby improving both reliability and precision.
2Reliability
If complex algorithms with or without associated lane markers are used, then detection completeness may be improved, but the algorithm complexity increases
Solution Approach 1:
The patent applies local quality by using different detection strategies in different regions: lane marker detection uses one approach while stop line detection uses another approach constrained by the lane marker geometry. This localized strategy improves detection completeness without requiring a single complex algorithm to handle all cases.
Solution Approach 2:
The patent performs preliminary detection of lane markers and establishes geometric constraints before conducting stop line detection. This preliminary action simplifies the subsequent stop line search by providing a constrained search space, reducing overall algorithm complexity while maintaining detection completeness.
3Productivity
If traditional detection systems are used, then the system structure is simple, but the processing time increases and vehicle safety is reduced
Solution Approach 1:
The patent segments the detection process into parallel independent stages: lane marker detection, candidate region identification, and stop line detection. This segmentation enables parallel processing of different detection tasks, improving processing speed while maintaining a relatively simple system structure.
Solution Approach 2:
The patent performs preliminary detection of lane markers and establishes geometric constraints that are reused in subsequent stop line detection. This preliminary action eliminates redundant computations, significantly reducing processing time while maintaining a simple overall system architecture.
Data Source
AI summary
A system and method for detecting a stop line on a roadway. The system and method a front view camera that generates images of the roadway. The system and method also include a controller that receives the images generated from the front view camera, including a bird's eye view image, said controller further programmed to provide a composite image that includes an original bird's eye view image and a rotated bird's eye view image. The controller is also programmed to use the composite image to determine if a stop line is present on the roadway.